Assessing the Style of Advance and Retreat of the Des Moines Lobe Using Lidar Topographic Data Sarah Elizabeth Day Iowa State University

Iowa State University Capstones, Theses and Graduate Theses and Dissertations Dissertations

2014 Assessing the style of advance and retreat of the Des Moines Lobe using LiDAR topographic data Sarah Elizabeth Day Iowa State University

Follow this and additional works at: https://lib.dr.iastate.edu/etd Part of the Geology Commons, and the Geomorphology Commons

Recommended Citation Day, Sarah Elizabeth, "Assessing the style of advance and retreat of the Des Moines Lobe using LiDAR topographic data" (2014). Graduate Theses and Dissertations. 13653. https://lib.dr.iastate.edu/etd/13653

This Thesis is brought to you for free and open access by the Iowa State University Capstones, Theses and Dissertations at Iowa State University Digital Repository. It has been accepted for inclusion in Graduate Theses and Dissertations by an authorized administrator of Iowa State University Digital Repository. For more information, please contact [email protected].

Assessing the style of advance and retreat of the Des Moines Lobe using LiDAR topographic data

Sarah Elizabeth Day

A thesis submitted to the graduate faculty

in partial fulfillment of the requirements for the degree of

MASTER OF SCIENCE

Major: Geology

Program of Study Committee: Neal R. Iverson, Major Professor Chris Harding William W. Simpkins

Iowa State University Ames, Iowa 2014

TABLE OF CONTENTS

Page

ACKNOWLEDGEMENTS ...... iii

ABSTRACT ...... iv

CHAPTER 1 INTRODUCTION ...... 1

A. The Des Moines Lobe ...... 1 B. Motivation and hypothesis ...... 15 C. Objectives ...... 16

CHAPTER 2 METHODS ...... 18

A. Introduction ...... 18 B. LiDAR ...... 19 C. Mapping methodology ...... 20 D. Testing retreat hypothesis ...... 29

CHAPTER 3 RESULTS ...... 32

A. Mapping ...... 32 B. Relationship of minor moraines to end moraines ...... 42

CHAPTER 4 DISCUSSION ...... 58

A. Evidence of stagnation ...... 58 B. Significance of minor moraine orientations ...... 62 C. Lack of flow-parallel features ...... 64 D. Modern analog ...... 67

CHAPTER 5 CONCLUSIONS ...... 70

REFERENCES ...... 72

iii

ACKNOWLEDGEMENTS

I would like to thank some of the people who helped me with this research and guided me through the writing process. First, I’d like to thank my advisor, Neal Iverson, for his guidance and patience throughout the entire research process. His insight and attention to detail have helped me make this thesis something I can be proud of.

Furthermore, given the additional responsibilities he took on as department chair, the amount of time he spent helping me with research and writing is astounding.

Next, I’d like to thank Lucas Zoet, Neal’s post-doc, who helped me appreciate the power of computer coding. He helped me process much of my data, and without his vast knowledge of different programing platforms, this research would not have been completed.

In addition, I would also like to thank a few other people who have helped me throughout this process. I’d like to thank the second and third members of my Program of Study committee, Chris Harding and Bill Simpkins, for their time and effort in perfecting this project. I would like to thank DeAnn Frisk, the department administrative specialist, for her patience, knowledge, and support. I would also like to thank the rest of the Iowa State geology department for making my time as a graduate student successful and enjoyable.

Finally, I’d like to thank my friends and family for supporting me even when I was frustrated. Without their love and support through the entire process I would not have been able to finish this journey as a graduate student.

ABSTRACT

Successive advances of the late-Wisconsinan Des Moines Lobe to form three major end moraines in Iowa—sequentially the Bemis, Altamont, and Algona moraines—are thought to be the result of the lobe surging out of balance with a warming climate.

Various styles of hummocky topography, collectively sometimes called stagnation moraine, are interpreted to be the result of widespread stagnation and down-wasting of ice following surges. Alternatively, end moraines could be recessional—a result of incremental back-wasting of the glacier margin and unrelated to surging.

To study the retreat style of the Des Moines Lobe, high resolution LiDAR data were used to re-evaluate the subtle landscape of the lobe’s footprint in Iowa. Results indicate that ~90% of the lobe’s area, excluding major Holocene stream drainages, consists of stagnation features. Some landforms are more prevalent than mapped previously, including eskers and features interpreted to be subdued ice-walled lake plains.

Importantly, subglacially formed minor moraines (a.k.a. washboard moraines), which resulted from sediment filling of transverse crevasses, cover ~60% of the lobe’s area with stagnation landforms. Also, ~25 previously unmapped end moraine ridges have been identified.

Transverse crevasse-fill ridges in the forefields of modern glaciers form due to longitudinal ice extension associated with surging and are not found in the forefields of non-surge-type glaciers, so minor moraines are good evidence of Des Moines Lobe surges. Most end moraines have minor moraine sets associated with them, consistent

with a surge-like advances, and many areas have multiple sets of minor moraines indicating a surge history more complicated than one advance for each of the three major end moraines. Therefore, asserting stagnation and down-wasting after three surge-like advances provides an incomplete characterization of the Des Moines Lobe’s advance and retreat. The surge-type Bering Glacier in Alaska is a good but imperfect modern analog for the lobe.

CHAPTER 1: INTRODUCTION

A. The Des Moines Lobe

The Des Moines Lobe, the largest of the lobes along the southern margin of the

Laurentide Ice Sheet (Fig. 1.1) during the last glacial maximum, advanced south into a relatively warm, boreal climate (Clark, 1994; Hooyer & Iverson, 2002; Mickelson &

Colgan, 2003). At its maximum extent at approximately 13,800 14C yr BP, the lobe covered more than 100,000 km2 in southern Minnesota and north-central Iowa and extended 600 km beyond the main body of the Laurentide Ice Sheet (Mathews, 1974;

Clayton & Moran, 1982; Patterson, 1997; Hooyer & Iverson, 2002). The lobe had retreated from Iowa by 11,700 14C yr BP (Clayton & Moran, 1982).

a b

Figure 1.1: a. Maximum extents of the Laurentide Ice Sheet and Cordilleran Ice Sheets during the late Wisconsinan and during earlier Quaternary glaciations (Illinois State Geological Survey, 2008). b. Extent of the southern Laurentide Ice Sheet lobes in the western Midwest ~13,800 radiocarbon years ago (Hooyer & Iverson, 2002).

A detailed chronology of the Des Moines Lobe in Iowa is well established by radiocarbon dates from trees overridden by the glacier and buried in till (Kemmis et al.

1981; Clayton & Moran, 1982). The Bemis (13,800 14C yr BP), Altamont (13,500 14C yr

BP), and Algona (12,300 14C yr BP) end moraines (Fig. 1.2) mark three major advances of the lobe (Ruhe, 1969; Kemmis et al., 1981). These moraines are broad, 1-10 km wide, swaths of hummocky topography characterized by linked-depressions, kames, ice- walled lake plains, and elongate flow parallel ridges (Kemmis, 1991).

Figure 1.2: Iowa footprint of the Des Moines Lobe. The three major end moraines are the Bemis, Altamont, and Algona, delimited by areas of hummocky topography. The moraine complex includes areas where the moraines merge laterally.

Paleoglaciology

Reconstructions of the Des Moines Lobe using modern moraine elevations and minor moraines to indicate flow direction indicate the lobe’s probable geometry at its maximum extent (Clark, 1992; Iverson and Hooyer, 2002). The lobe was unusually thin

and gently sloping. Clark (1992) calculated its thickness to be only 300 m, 275 km up- glacier from the terminus, a factor of ~10 thinner than the margin of the Greenland Ice

Sheet. Hooyer and Iverson (2002), assuming the moraines may have been ice cored at the time of deposition, determined only slightly larger thicknesses. These estimates are consistent with the earlier reconstruction of Matthews (1974), who did not use minor moraines as flow direction indicators.

Driving stresses computed from reconstructions range from 0.7 to 15 kPa (Clark,

1992; Hooyer & Iverson, 2002), about 3-70 times lower than those of a typical glacier,

~50 kPa (Cuffey & Paterson, 2010). Owing to the lack of ice bordering the lateral or terminal margins of the Des Moines Lobe and its large area relative to its thickness, unlike many modern ice streams (Alley et al., 1987) the reconstructed driving stress of the lobe should provide a good estimate of its average basal shear stress, at least in areas away from the glacier margin where slopes do not vary too much spatially. The low basal shear stress indicates that the lobe was likely not frozen to its bed, an inference consistent with the relatively warm climate inferred from fossil vegetation and insects

(Schwert & Torpen, 1996; Baker, 1996; Bettis et al., 1996).

Even though the Des Moines Lobe was thin, gently sloping and had low driving stresses, it had a high ice flow velocity. Radiocarbon dates of wood found within end moraines and farther up-glacier indicate that the ice flowed rapidly to each end moraine position (Clayton et al., 1985). The time-averaged rate of margin advance of the Des

Moines Lobe has been estimated to be 2 km yr-1, based on radiocarbon dates along flow lines (Clayton et al., 1985). This value is much higher than the range of measured

velocities for non-surge-type mountain glaciers and many ice-sheet outlet glaciers, approximately 10-200 m yr-1, but is comparable to speeds of some surge-type glaciers, tidewater glaciers, and ice streams (Cuffey & Paterson, 2010).

The Des Moines Lobe moved over a fine-grained basal till. It is generally massive and relatively homogeneous, made up of ~15% clay, ~36% silt, and ~47% sand and larger particles (Kemmis et al., 1981), although it contains isolated pods and stringers of sand and gravel. The till’s fine-grained texture, together with the low slope of the lobe would have, according to the leading theoretical model of subglacial water flow for soft- bedded glaciers (Clark & Walder, 1994; Walder & Fowler, 1994), resulted in channels cut into the bed surface and under high water pressure, resulting in high pore-water pressure in adjacent till. This conjecture is consistent with the sparseness of well develop eskers within the limits of the Des Moines Lobe, which require R-channels cut upward into the glacier sole (Clark & Walder, 1994). High basal pore-water pressure would have reduced the till’s shear strength and could have resulted in basal till deformation that promoted unstable and rapid glacier flow (Clark, 1994). Clark (1994) found that areas of the Laurentide ice sheet that experienced unstable rapid ice-margin fluctuations (Fig.

1.3a) were underlain by fine-grained till and these regions tend to correspond to areas of sedimentary bedrock (Fig. 1.3b). He argued that till deformation occurred to depths as great as 10 m. Many have noted that high basal pore water pressure and resultant till deformation can produce subglacial topography (Alley, 1991; Clark, 1994; Eyles et al.,

1999; Boone & Eyles, 2001), for example, where sediment is squeezed upward into

crevasses at the bed, forming crevasse-fill ridges (e.g., Sharp, 1985a; Rea and Evans,

2011).

a b

Figure 1.3: a. The time-distance plot for the Des Moines Lobe constructed using radiocarbon dates of wood entrained in the till. b. Generalized map with the distribution of crystalline bedrock (patterned area) and sedimentary bedrock (unpatterned area) (Clark, 1994; Clayton & Moran, 1982).

Hooyer and Iverson (2002) did preconsolidation testing on intact basal till samples to help confirm that high basal pore water pressure may have been responsible for the low basal shear stress of the lobe and its seemingly high velocity. The maximum past effective stress on the till determined from these tests, when considered together with the lobe’s thickness determined from the geomorphic reconstruction, allowed the minimum basal pore water pressure to be determined (Hooyer &Iverson, 2002). The corresponding potentiometric surface plotted on surface profiles of the Des Moines Lobe

(Fig. 1.4) indicates that pore-water pressures were sufficiently high that the lobe was nearly floating on its bed (Hooyer & Iverson, 2002). Following the model of Walder and Fowler (1994), the subglacial hydrology may well have consisted of a distributed system with channels cut into the bed supporting higher water pressures (Hooyer &

Iverson, 2002). Unlike Clark (1994), relatively weak fabrics in the basal till of the lobe

a b

Figure 1.4: a. Longitudinal and b. transverse potentiometric surfaces of the Des Moines Lobe. Two ice surface profiles and two potentiometric surfaces are depicted in a. and b.; the lower of the two ice surface profiles was calculated assuming that the elevation of the Bemis moraine reflects the elevation of the ice surface at the margin. The higher ice surface profile assumes that Bemis moraine was ice cored during deposition and hence at a higher elevation than it is today. A potentiometric surface was calculated for each of these ice surfaces using both preconsolidation pressures measured from intact samples of the Des Moines Lobe basal till and ice pressures based on the reconstructed ice thickness (Hooyer & Iverson, 2002).

(the Alden member of the Dows Formation (Kemmis et al., 1981; Bettis et al., 1996)) measured by Hooyer and Iverson (2002) led them to conclude that the glacier was likely decoupled from its bed over much of its area, with clast plowing as the dominant mechanism of motion rather than pervasive and deep deformation of till (Alley, 1989;

Clark, 1994; Boulton, 1996).

Models of Des Moines Lobe motion

Early workers believed that the Des Moines Lobe underwent climate driven advance and retreat in response to changes in the glacier’s mass balance. End moraines and minor moraines, also known as washboard moraines (Elson, 1968; Cline, 2011) and corrugated moraines (Prest, 1968; Stewart et al., 1988), were cited as evidence (Gwynne,

1942; Gwynne, 1951; Lawrence & Elson, 1953). Gwynne (1942, 1951) and Elson

(1953) were well aware that end moraines outline positions of the ice front and formed in locations where the ice margin was temporarily stable. They considered minor moraines to also outline successive positions of the ice front but where the ice margin paused for a shorter period than at the end moraines (Gwynne, 1942; Gwynne, 1951;

Lawrence & Elson, 1953). Each minor moraine was estimated to represent approximately one to two years of deposition at the ice margin (Gwynne, 1942; Gwynne,

1951; Lawrence & Elson, 1953). Gwynne (1942) calculated the retreat rate of the lobe by counting the number of minor moraines along a segment of a flow line and then calculating the rate as one to two times the number of minor moraines divided by the distance. Areas of the Des Moines Lobe that do not have minor moraines were considered to be regions where the ice margin retreated too rapidly to deposit a moraine

(Gwynne, 1942; Gwynne, 1951; Lawrence & Elson, 1953).

The idea that the Des Moines Lobe advanced and retreated in response to changes in its mass balance was superseded in the 1980s by the hypothesis that the lobe advanced by surging. A surge-type glacier is characterized by periodic rapid flow events, or surges, in which ice velocity increases by a factor of 10 to 100 (Cuffey & Paterson,

2010). High ice velocity, uncompensated by an associated increase in glacier net balance, results in a reduction of ice surface slope and thickness followed by widespread reduction in flow velocity that transiently results in nearly complete ice stagnation. The

Des Moines Lobe’s low surface slope, thin ice, and advance rate of 2 km yr-1 are consistent with the characteristics of surging glaciers (Clayton et al., 1985). Subsequent surge-type advances occur only after a period of many years, commonly 10-100 years in

valley glaciers (Cuffey & Paterson, 2010),

once the glacier surface slope and

thickness have recovered to threshold

levels necessary for another surge (Kamb

et al., 1985; Roush et al., 2003). For surge-

type glacier advances, each end moraine

Figure 1.5: Stagnation and down represents the farthest down-glacier extent wasting of the margin of a surge-type glacier, resulting in the formation of of a given surge front (Fig. 1.5). hummocky topography (Hooyer & Iverson, 2002). Landforms such as minor moraines and ice-walled lake plains are thought to develop after a surge during stagnation (Sharp,

1985a; Boone & Eyles, 2001; Clayton et al., 2008; Rea & Evans, 2011).

The Des Moines Lobe’s low driving stresses, high basal pore water pressure, and gently sloping profile are also consistent with some of the characteristics of ice streams

(Fig. 1.6) (Patterson, 1997; Patterson, 1998; Jennings, 2006). Ice streams are zones of rapidly moving ice, 300-15,000 m yr-1 (Cuffey & Paterson, 2010), flanked by slowly moving ice (Clarke, 1987; Stokes & Clark, 1999). As noted, the rapid advance the Des

Moines Lobe is well established based on radiocarbon dating. There is less evidence adjacent to the lobe for the slowly moving ice which flanks an ice stream. Over the lobe’s full extent in Iowa there was no ice along the lobe’s lateral flanks. Farther north there may have been slower moving adjacent ice as a result of escarpments and bedrock topographic features that restricted ice flow (Patterson, 1997; Patterson, 1998; Stokes &

Clark, 1999; Stokes & Clark, 2001; Jennings, 2006).

Patterson (1997) discussed landforms of the

Des Moines Lobe that may indicate it was an ice

stream, but these characteristics are not unique to

ice streams and more characteristic of surge-type

glaciers. According to Patterson (1997) tunnel

valleys at the glacier margin would have increased

basal drag by evacuating large amounts of water

from the bed, thereby causing stagnation. Although

the evacuation of water through tunnel valleys may

be associated with ice streams, this style of water

evacuation has been linked more commonly to the

Figure 1.6: Map of the Des periodic advances of modern surging glaciers Moines Lobe, considered to be an ice stream by Patterson (Kamb et al., 1985; Walder & Fowler, 1994; (1997). Roush et al., 2003), with high discharges occurring as the surge ends. Patterson (1997) also links transverse crevasses and the development of minor moraines to ice stream motion. Transverse crevasses develop due to longitudinal ice extension, (Sharp, 1985a) and can result in crevasse fill ridges, interpreted by many to be the origin of minor moraines (Stewart et al., 1988; Clark,

1992; Patterson, 1998; Ankerstjerne, 2010; Cline, 2011). Crevasse-fill ridges are associated most commonly with the forefields of surge-type glaciers rather than ice streams (Sharp, 1985a; Rea and Evans, 2011).

Landform identification

Landforms of the Des Moines Lobe have been studied in Iowa for more than 100 years (Chamberlin, 1883; Macbride, T.H., 1909), but the most detailed Quaternary landform map was produced by Kemmis (1991). Kemmis’s map (Fig. 1.7) covers the northernmost third of the Des Moines Lobe in Iowa. Using low-resolution aerial photographs and field studies, Kemmis (1981, 1991) developed a complex landform classification scheme (Fig. 1.8) with 24 different positive-relief features, such as moraines and plains, and additional negative-relief features including abandoned channels and linked depressions (Kemmis, 1991).

His positive-relief landforms include ridges and plains. Hummocky moraine ridges are defined as having a distinct front and back slope. These moraines are 1-10 km in width and are locally interrupted by stream valleys or abandoned channels (Kemmis,

1991; Bettis et al., 1996). Plains are broad extensive areas of moderate to low relief and have various surface patterns (Bettis et al., 1996); they are the most diverse and extensive landform identified. Transverse lineated low relief features or minor moraines are one such surface pattern. A second surface pattern was identified based on rounded circular features (Kemmis, 1991; Bettis et al., 1996), possible ice-walled lake plains

(Fig. 1.9), which form when a glacial lake develops within the ice and subsequent melting leaves a topographically inverted, circular, flat-topped plain (Clayton et al.,

2008).

Figure 1.7: Quaternary landform map of the northern third of the Des Moines Lobe in Iowa (from Kemmis 1991).

Escarpments Ramp-like Slopes—NRR1 Plains Hummocky Plateau—HP1 Ridge Plateau-like Plains

Hummocky Non Hummocky w/ Lineated Surface Patterns PR1 PU1 PS1 PD1 PH1 Hummocky Non Hummocky PP1 PP2 PP3 HE1 HE2 NE1 NE2 NE3 PL1 PL2 PL3 PL4 HRS1 HRS2 HRS3 NRS1 NRS2

Figure 1.8: Landform classification scheme of Kemmis (1993). See legend of Figure 1.7 for descriptions of landform types (e.g. PL3).

Figure 1.9: Ice-walled lake plain formation a. prior to ice fully melting and b. following ice melting (Clayton et. al, 2008).

Negative-relief landforms are abandoned channels and linked depressions (Bettis et al., 1996). Abandoned channels are former tunnel valleys (fig. 10) or more minor drainages and are frequently connected to or are part of a linked depression system

(Kemmis, 1991; Bettis et al., 1996). Linked depression systems are valleys and depressions that connect to form drainage networks, often found super-imposed on positive-relief landforms (Kemmis 1991). They tend not to have smoothly graded

channel bottoms, are branching rather than linear, and typically connect to modern drainages (Kemmis, 1991; Bettis et al., 1996).

b c

Figure 1.10: a. Tunnel valley south of Mallard. b. Clear Lake, Iowa, inferred to occupy a tunnel valley. c. Map of Des Moines Lobe extent of tunnel valleys (Patterson, 1997). Note that no tunnel valleys have been previously mapped south of the Kemmis map area.

Although Kemmis (1991) developed a very detailed map and classification scheme, his work is insufficient to describe the lobe in Iowa. He mapped only the northern third of the lobe in Iowa (Fig. 1.7), and when the map was completed, high resolution 1 m

DEMs were not yet available. Due to the subtle relief over the footprint of the lobe, high resolution imagery can illuminate previously overlooked landforms. Additionally,

Kemmis’ classification scheme (Fig. 1.8) is very complex, including 24 landforms, nine of which are different types of plains. The distinction between the nine types of plain identified by Kemmis is not clear; some definitions overlap others (Fig. 1.7). In order to

gain a better understanding of the lobe in Iowa, landforms over the entire footprint of the lobe in Iowa need to be examined with the newly available 1 m DEMs and the Kemmis classification scheme re-evaluated.

Additional landforms have also been examined separate from Kemmis’ (1991) work such as doubly blown-out doughnuts and eskers. Although doubly blown-out doughnuts have not previously been mapped in Iowa, they have been identified in the glaciated landscape of North Dakota as part of the Northern James River and Des Moines Lobes

(Fig. 1. 1). Doughnuts, before being truncated by erosion (blown-out), tend to be circular ridges a few hundred feet across and a few tens of feet high (Clayton & Freers,

1967). These features form in three stages (Fig. 1.11). First, sediment fills a depression in stagnant ice. Second, the surrounding ice melts, leaving an ice cored hill, and third,

Figure 1.11: Three stages of doughnut formation (Clayton, 1967). the ice core melts and a depression forms at the center, leaving a circular ridge (Clayton

& Freers, 1967). Doughnuts are commonly breached on opposite sides of the ridge ring, possibly resulting from the collapse of a buried ice-cored crevasse fill (Clayton & Freers,

1967), and identified as doubly blown-out doughnuts. Eskers have also been identified in the footprint of the lobe. Traditionally, very few eskers have been identified in Iowa, and are located near the Iowa-Minnesota border (Quade & Seigley, 2006). Additional sinuous ridges, that are similar to the morphology of an esker, have been noted but not

previously identified as eskers. One such sinuous ridge is located ~ 3 km north of Lake

Mills, Iowa, (Fig. 1.12) identified by Leverett in 1932 as a moraine ridge.

Figure 1.12: Esker north of Lake Mills, Iowa, near the Minnesota border. The esker is approximately 5 m high at its crest and extends to eastern margin of the lobe.

B. Motivation and hypotheses

Identification and interpretation of glacial landforms can improve reconstructions of ancient ice masses. Landforms provide information about glacier morphology and dynamics. In previous reconstructions of the Des Moines Lobe, the assumption is made

that landforms such as minor moraines all formed simultaneously when the glacier was at its maximum extent (Clark, 1992; Hooyer & Iverson, 2002). The veracity of these reconstructions can be evaluated by developing more complete, high resolution maps of the lobe’s end and minor moraines.

Landforms of the lobe also provide information about its style of advance and retreat. As noted, Gwynne (1942) proposed that the Des Moines Lobe underwent advance and retreat due to mass balance changes, which would indicate that radiocarbon dating of moraines would provide information about past climate change. In contrast, others have subsequently suggested (Clark, 1994; Clayton et al., 1985; Patterson, 1997) that the Des Moines Lobe advanced out of balance with the climate. Of these hypotheses, the most popular is that the lobe underwent surge-type advances followed by stagnation. If surges were the dominant style of advance, it is important to understand how many advances there were. Traditionally, the Des Moines Lobe is thought to have undergone three surge-like advances, each building one of the three major end moraines

(Fig. 1.2). A more detailed analysis of Des Moines Lobe landforms may support this hypothesis, reveal that the that the surge history of the lobe is more complex than previously thought, or provide evidence for some moraines forming during active recession of the lobe, rather than by surging and down-wasting.

C. Objectives

The goal of this project was to use high-resolution LiDAR-derived elevation data to map the Des Moines Lobe footprint in Iowa and thereby better assess the advance and

retreat style of the lobe using landform attributes. Landforms over the 30,000 km2 footprint of the lobe in Iowa were identified, including minor moraines (Stewart et al.,

1988), end moraines (Kemmis, 1991; Benn, 1992), ice-walled lake plains (Clayton et al.,

2008), doubly blown-out doughnuts (Clayton & Freers, 1967), eskers (Shreve, 1985), and tunnel valleys (Kemmis, 1991; Patterson, 1997). New maps were used to identify the total area that is characterized by landforms that developed as a result of stagnation

(Sharp, 1985a; Benn, 1992; Clayton et al., 2008) and to determine the relationship between orientations of end moraines and minor moraines—a key relationship for assessing the extent to which surge-like advances and stagnation dominated the advance and retreat behavior of the lobe.

CHAPTER 2: METHODS

A. Introduction

Previous work on the landscape of the Des Moines Lobe, completed by Kemmis in

1991, used aerial photography, field observations, and laboratory work to map the northern third of the Des Moines Lobe’s footprint in Iowa (Fig. 2.1). Kemmis developed

Figure 2.1: Kemmis (1991) study area, outlined in red, overlaying 1 m DEMs derived from LiDAR data and a hill-shaded relief map. a new landform classification scheme (Fig. 1.8), and although his map is detailed, most of the Des Moines Lobe’s footprint in Iowa is not included. Moreover, he was not able to observe the extent of many of the landforms and overlooked smaller, more subtle features. The limitations of this previous work and the availability of new high resolution imaging over the entire state of Iowa motivated this project. Initial

observations of LiDAR derived DEMs highlighted a number of areas of interest that appeared to differ from or expand upon Kemmis’s results. New maps covering the full extent of the lobe in Iowa allowed advance and retreat hypotheses for the lobe, studied by Kemmis (1991), to be reexamined.

This project used LiDAR (Light Detection and Ranging) data to produce a one-meter digital elevation model (DEM) highlighting landforms. The landforms identified include end moraines, minor moraines, ice-walled lake plains, doubly breached doughnuts, eskers, and tunnel valleys. Once they were identified and mapped, the landforms were grouped into a simplified landform regions map.

B. LiDAR

LiDAR-derived elevation data are accurate to the sub-meter scale and are particularly useful for the low-relief landscape of the Des Moines Lobe. The DNR’s

Iowa LiDAR Mapping Project (Iowa Department of Natural Resources) used LiDAR elevation data, collected from 2006 through 2009, to generate high-resolution one meter

DEMs for the state of Iowa (Cline, 2011). With this technique, elevation is detected using light pulses; millions of pulses per second are emitted and reflected back to a single aircraft (Measures, 1984; Elachi & van Zyl, 2006). The pulse lag time and GPS position of the aircraft is used to calculate absolute elevations. Each pulse is emitted with a wavelength of 1064 nm, a much shorter wavelength than the radio waves used by radar (approximately 0.01 – 1 m wavelength) (Measures, 1984; Cracknell & Hayes,

1991; Elachi & van Zyl, 2006). As a result, LiDAR data are much higher resolution.

LiDAR data reveal details not visible on aerial photographs or 10 m DEMs (Fig.

2.2). Figure 2.2 depicts an area between Ames and Nevada, Iowa where minor moraines are prevalent. The 10 m DEM (Fig. 2.2a) has visible minor moraines over only approximately 2 km2 of the map area, whereas in the 1 m LiDAR based DEM (Fig. 2.2b) most of the approximately 16 km2 depicted is covered by well-resolved minor moraines.

This study is the first to use LiDAR elevation data to examine the Des Moines Lobe landscape of Iowa and expands upon Cline’s (2011) work on the washboard moraines of

Story County.

Figure 2.2: a. Map showing a 10 m DEM of the area between Ames and Nevada, Iowa, as was available prior to the LiDAR survey (Cline, 2011). b. LiDAR-derived DEM (1 m horizontal resolution with vertical accuracy within 18.5 cm) of the same area (Cline 2011). Both parts a and b are overlain with a hill-shaded relief map to highlight topography.

C. Mapping methodology

A landform map was generated by categorizing topographic variability over the footprint of the Des Moines Lobe. Initial landform mapping was completed by

identifying individual landforms based on their size and shape using the one-meter

DEMs developed from the LiDAR elevation data. Some of the landform classifications were optimized through field checking of a few areas of interest, as well as by using previous field work (Kemmis, 1991; Ankerstjerne, 2010; Cline, 2011). A landform classification scheme was developed to reflect a simplified version of Kemmis’s (1991) scheme of 24 landform groups and improved through use of modern definitions and understanding of glacial landforms (Fig. 2.3, Table 2.1).

Figure 2.3: Classification scheme for the landform regions map.

End moraines are defined as either broad hummocky moraines or moraine ridges.

Broad hummocky end moraines, such as the Altamont moraine (Fig. 2.4a), have approximate heights of 5 to 15 m, are 1 to 10 km wide, and are characterized by hummocky topography that may contain multiple ridges (Kemmis, 1991). In contrast, end moraine ridges have one distinct ridge form with a foreslope and a backslope and are generally 5-10 m high, with a width of ~1 km. (Fig. 2.4b, c).

Additional landforms identified, including minor moraines, ice-walled lake plains, and doubly blown-out doughnuts, occur in groups across broad plains. Minor moraines

(Fig 2.4d, e), are composed of subglacial till with minor sand and gravel. Each moraine has a distinct ridge form with amplitude of 1-5 m and wavelength of ~100 m (Cline,

Table 2.1: Descriptions of the landform regions. Landform Region Description End moraine Distinct ridge form or topographically high broad swath of hummocky topography that may contain distinct ridges.

Plain with circular features Broad area of low relief containing depressions and hills that are quasi- circular in map view. Some of the hill features are characterized by relatively flat surfaces indicative of ice-walled lake plains.

Plain with doubly blown out doughnuts Broad area of low relief with pairs of short parallel ridges, approximately 100 m long, one to two meters high, and 30- 50 m apart.

Plain with minor moraines Broad area of low relief with minor moraine ridges. There may be more than one set of minor moraines with varying trends.

Hybrid Plain with minor moraines and Broad area of low relief with a circular features combination of minor moraines and quasi-circular features.

Hybrid Plain with minor moraines, Broad area of low relief with a circular features and doubly blown out combination of minor moraines, circular doughnuts features, and doubly blown-out doughnuts.

Hybrid Plain with circular features and Broad area of low relief with a doubly blown out doughnuts combination of circular features and doubly blown out doughnuts.

Till Plain Broad area of low relief with no distinct landforms.

2011). Ice-walled lake plains (Fig. 2.4f) are characterized by a quasi-circular, flat- topped hill composed of sorted sand. Doubly blown-out doughnuts are breached

(commonly on opposite sides) circular ridges (Clayton & Freers, 1967), resulting in short (100 m or less), paired, parallel ridges with heights of 1-2 m and spacings between ridges of 20-50 m (Fig. 2.4g).

Channelized subglacial drainage of water resulted in two other less prevalent landforms: eskers and tunnel valleys. Eskers are sinuous ridges perpendicular to the ice margin composed of sorted sand and gravel (Fig 1.12). They likely represent locations where a subglacial tunnel carried water toward the margin and the tunnel filled with sediment. Alternatively, eskers could reflect supraglacial channels that contained sediment that was let down on the landscape during stagnation (Benn and Evans, 2010).

Tunnel valleys (Fig. 1.10), also typically perpendicular to the margin, are valleys where water flowed at the glacier bed near the margin. Water likely did not fill the valley but occupied a smaller channel that migrated back and forth across the bed to form the valley (Cofaigh, 1996; Clayton et. al, 1999).

Eight landform regions are divided into two major types: moraines and plains. End moraines, as noted, are characterized either by broad belts of hummocky topography, such as the Altamont Moraine (Fig. 2.4a) (Kemmis, 1991), or by one distinct ridge form with a fore-slope and a back-slope (Fig. 2.4b, c). The plain regions are subdivided based on whether minor moraines, circular features such as ice-walled lake plains and kames, doubly

b c

Figure 2.4: Examples of each of the eight landform regions. The Altamont moraine a., a broad hummocky end moraine, delimited by black lines. The area shown is just north of Boone, Iowa, and east of Fraser, Iowa. Note the multiple ridges and undulating topography within the area of the broad ridge. b. End-moraine ridges south of the Algona moraine. c. A closer view of the same end moraines south of the Algona moraine and east of Humboldt, Iowa. Plains with minor moraines, d. trending north-east between Ames, Iowa and Nevada, Iowa, and e. with multiple sets southeast of Gilmore City, Iowa. f. Plain with circular features, ice walled lake plains, west of Dolliver, Iowa. g. Plain with doubly blown-out doughnuts east of Forest City, Iowa. Note the short ridge pairs in the central portion of the map. h. Plain with circular features and minor moraines east of Dolliver, Iowa. i. Plain with circular features, minor moraines, and doubly blown-out doughnuts south of Clear Lake, Iowa. j. Plain with circular features and doubly blown-out doughnuts west of Joice, Iowa. k. Till plain south of the Algona moraine near Algona, Iowa.

Figure 2.4 continued:

blown-out doughnuts, or a combination of these are present. Hybrid plains are defined as a plain with a combination of landforms. There are three sub-categories of hybrid plains: those with circular features and minor moraines, with circular features and doubly blown out doughnuts, and one with all three landform types. Additionally, there are small plains where there is no evidence of any specific landforms; inferring they consist of basal till, they are called till plains, (Fig. 2.3, Table 2.1). Examples of each of the eight landform regions are shown in Fig. 2.4 a-k.

D. Testing retreat hypothesis

One purpose of the maps was to examine the relationship between minor moraines and end moraines to try to test different retreat hypotheses for the Des Moines Lobe. As noted in Chapter 1, the leading hypothesis is that the lobe surged three times out of balance with a warming climate, then stagnated and down-wasted to form the three major end moraines. Alternatively, the lobe could have receded as a result of incremental back-wasting of the margin, unrelated to surging, leaving recessional end moraines. Diverse evidence indicates that minor moraines form as subglacial crevasse fills (Stewart et al., 1988; Ankerstjerne, 2010; Cline, 2011). During a surge the glacier undergoes longitudinal extension, resulting in the development of rather uniformly spaced transverse basal crevasses. Subglacial till then deforms into basal crevasses during stagnation leaving minor moraines (Stewart et al., 1988; Ankerstjerne, 2010;

Cline, 2011) with amplitudes that degrade with subsequent hill-slope processes (Burras

& Scholtes, 1987). Over much of the lobe’s area there are locations where there are two

or more orientations of minor moraines. If the leading hypothesis is true, such that end moraines formed by surging, then minor moraine orientations up-glacier from each end moraine should parallel the down-glacier end moraine. In contrast, if an end moraine has no parallel minor moraine orientations, it could reflect active glacier recession.

In order to study minor moraine orientations, flow paths were defined based on the reconstruction of the lobe at its maximum extent by Hooyer and Iverson (2002). Using

MATLAB scripts to automate the process, positions were picked every 5 km along 19 flow paths north from the Bemis moraine margin to the Minnesota border. A box, 5 km2, was drawn around each position, and the orientation for every minor moraine within the sample area was measured. The orientation of each measured moraine was then calculated and sorted into five degree bins for a rose diagram. Of the 19 flow paths used, 573 rose diagrams were produced. These diagrams were then used to illustrate the orientations of the minor moraines at each position.

One to three dominant minor moraine orientations at each location were compared with the orientations of the nearest down glacier end moraine. Angular differences between minor moraines that are most parallel and least parallel to the orientations of the nearest down-glacier end moraine were measured. Averages and standard deviations of these angular differences were calculated by considering only the sampled area immediately up-glacier from the moraine along multiple flow paths and by also considering all sampled areas between the end moraine immediately down-glacier and the next end moraine up-glacier. This process helped illuminate whether end moraines

had associated (parallel or sub-parallel) minor moraine sets—the criterion used in this thesis for a surge-type end moraine.

CHAPTER 3: RESULTS

A. Mapping

Quaternary landform mapping of the Des Moines Lobe footprint in Iowa resulted in a map of major landforms (Fig. 3.1a) and second map of landform groups (Fig. 3.1b), each covering approximately 30,000 square kilometers. Previous Quaternary maps of the area included broad areas of hummocky topography, also known as stagnation moraine

(Hallberg et al., 1991), including the three major end moraines of the lobe: the Bemis,

Altamont, and Algona (Fig. 1.2). These maps also include moraine complexes where these moraines merge, intervening areas of low-relief “ground moraine,” and three small end moraines (the Clare, Renwick, and West Bend). The new maps of this thesis provide a more detailed picture of the Des Moines landscape.

The new Quaternary landform map (Fig. 3.1a) expands upon Kemmis’ work (1991) on the northernmost third of the lobe’s extent in Iowa (Fig. 1.8), with a greater diversity of landforms identified in intra-moraine areas previously identified as stagnation or ground moraine (Hallberg et al., 1991). Landforms that are more prevalent than previously mapped include minor moraines (Stewart et al., 1988), ice-walled lake plains

(Clayton et al., 2008), and eskers (Shreve, 1985). Doubly blown-out doughnuts (Clayton

& Freers, 1967), previously unmapped for the Des Moines Lobe, were identified. In addition to stagnation landforms, approximately 25 previously unmapped end moraines were mapped.

Drainages

Figure 3.1: a. Quaternary landform map of the Des Moines Lobe landscape in Iowa with individual landforms overlaying 1 m resolution DEMs derived from LiDAR data and with a hill-shaded relief map. b. Landform regions map, with stream drainages superimposed on landform regions. Full size copies of these maps are located in the supplementary information.

The new landform regions map (Fig. 3.1b) indicates that 90% of the lobe’s Iowa extent, without considering modern drainage systems, is characterized by stagnation landforms. The most important of these landforms are minor moraines, found across plains of minor moraines and hybrid plains, which comprise 60% of the area covered by stagnation features. The rest of the area covered by stagnation features includes hummocky topography (31%) common in the three major end moraine areas, ice-walled lake plains (6%), and doubly blown-out doughnuts (3%).

Minor moraines

One-meter LiDAR-based DEMs allowed both well-developed (Fig. 3.2a) and more subtle (Fig. 3.2b, c) minor moraines to be identified across the lobe (Fig. 3.2). Although

b c

Figure 3.2: 1 m DEMs of a. Well-developed minor moraines, trending southwest, immediately east of Ames, Iowa. b. More subtle minor moraines, with less lateral continuity east of Churdan, Iowa. c. The minor moraines of b. with their trends outlined.

some minor moraines were difficult to identify due to limited continuity of their crests, any ridge set that could be traced as parallel or sub-parallel discontinuous lineaments were mapped as minor moraines. Minor moraines are generally 1-5 m in amplitude with a wavelength of approximately 100 m (Fig. 3.3). For the well-defined minor moraines of

301 a

300

299 331 b 298 330

297 329

Figure 3.3: Topographic profile from a. Fig 3.2a and b. Fig. 3.2b marked by solid black line.

Story County, Iowa—the focus of Cline’s (2011) study of minor moraine spatial attributes—Fourier analysis yielded a dominant ridge crest spacing of 110 m, with a standard deviation of 45.4 m. Well-defined minor moraines tend to be focused near the three major end moraines and tend to become more subtle and discontinuous farther from them. Also, away from the three major end moraines, minor moraines occasionally are well-developed, like those between the Renwick and West Bend moraines (Fig.

3.2a). Minor moraines are absent between the Algona and West Bend end moraines, as well in parts of the Bemis-Altamont moraine complex (Ruhe, 1952; Quade et al., 2005;

Cline, 2011).

Across much of the lobe, minor moraines occur in two to three different sets, each with a distinct orientation (Fig. 3.4). The number of minor moraine sets varies throughout the lobe. Near the Bemis moraine there is generally only one set of minor

moraines. Farther up-glacier there tends to be two to three sets. Not uncommonly, the number of minor moraine sets is different immediately up-glacier or down-glacier from an end moraine.

a b

Figure 3.4: Area containing two minor sets. a. Map of area southwest of Swea City, Iowa, and north of the Algona moraine, with moraine trends outlined in the lower image, and b. of the area surrounding Buckeye, Iowa, southeast of the Altamont moraine.

In some areas of the lobe, minor moraines are associated with other stagnation features such as ice-walled lake plains and doubly blown-out doughnuts. Minor moraines occur together with these landforms most commonly near the three major end moraines. Minor moraines can be superimposed on ice-walled lake plains (Fig. 3.5). In

Figure 3.5: Minor moraines superimposed on ice-walled lake plains north of Ringsted, Iowa. this case, the moraines are discontinuous in the areas between the ice-walled lake plains but can be traced across their flat interior surfaces. On these surfaces, minor moraines tend to have one dominant orientation, but in other regions of the lobe, as noted, multiple sets of minor moraines with different orientations are commonly present.

End moraines

Approximately 25 previously unmapped end moraine ridges were identified that are not apparent in lower resolution imagery, in addition to the broad belts of hummocky

365 b 360 355 West Bend 3 50 EM8 EM9

0 1 2 3 4 5 6 7 8 9 10 11

375 Figure 3.6: a. Map of the end c moraines identified within the Des 370 Moines Lobe. b. Topographic profile EM2 365 of West Bend moraine and of newly

360 mapped end moraines EM8 and EM9. 0 1 2 3 4 c. Topographic profile of EM2.

topography that constitute the Bemis, Altamont, Algona, and West Bend moraines and the narrow ridges with more distinct crests that constitute the Renwick and Clare moraines. The newly mapped ridges are lobate, with wavelengths of approximately 1-3 km and amplitudes of approximately 5-10 m (Fig 2.4 & Fig. 3.6).

Additional stagnation features

Doubly blown-out doughnuts (Clayton & Freers, 1967) are a landform that has not been identified in previous mapping efforts in Iowa (Fig. 3.7). These landforms are typically expressed as pairs of short parallel ridges that occur in groups with multiple orientations (Fig. 3.7a), unsystematically related to the glacier flow direction. Most ridges are 100 m or less in length, with heights of 1-2 m. The ridges of a given pair are spaced 20-50 m apart. These features are most common in the Bemis-Altamont moraine complex in the northern half of the map area (Fig. 3.7b).

a b

Figure 3.7: a. Doubly-breached doughnuts near Pilot Knob State Park east of Forest City, Iowa. b. Areal extent of doubly-breached doughnuts (green). Black lines show the distal extent of the three major end moraines.

Similar to minor moraines and end-moraine ridges, the extent and number of ice- walled lake plains are greater than previously mapped. They are more subtle than those described for more northerly lobes of the Laurentide ice sheet (Clayton et al., 2008), with heights of approximately 1-5 m and irregularly shaped areas of approximately 750 m2 to 2.5 km2. They occur most commonly in clusters in the northern portion of the map area and near the three major end moraines (Fig. 3.8). In addition to commonly being associated with minor moraines, ice-walled lake plains can be spatially associated with doubly blown-out doughnuts.

Figure 3.8: Map of areal extent of ice walled lake plains within the Des Moines Lobe footprint (blue). Black lines show the distal extent of the three major end moraines.

Eskers and tunnel valleys

Eskers and tunnel valleys have been extensively mapped in the northern portion of the Des Moines Lobe (Kemmis, 1991; Patterson, 1997). Four additional eskers and five

tunnel valleys were identified and mapped in this study, largely on the basis of their morphology (Fig. 3.9). Eskers are conspicuously sinuous ridges that are generally perpendicular to the margin. Those mapped tend to have lengths of 1-20 km and heights of 0.5-5 m. Tunnel valleys (e.g., Benn & Evens, 2010), relatively linear channel segments of length 1-8 km that are inset into the till upland 2-15 m, are also generally perpendicular to the margin and either lack a modern stream drainage or contain an elongate lake (Fig. 1.10). The eskers and tunnel valleys are located most commonly within the Bemis-Altamont moraine complex and the three major end moraines (Fig.

3.9). A few of the newly mapped eskers and tunnel valleys are adjacent and

Figure 3.9: Eskers, including four newly identified, and tunnel valleys. Note only tunnel valleys without modern drainages have been included, due to the uncertainty of identifying tunnel valleys that contain a modern drainage.

perpendicular to some of the newly mapped end moraines between the Altamont and

Algona moraines. Previously mapped eskers occur only in the northern portion of the map area, near the Iowa-Minnesota border. No tunnel valleys have previously been mapped south of the area mapped by Kemmis in 1991 (Fig. 1.10c).

B. Relationship of minor moraines to end moraines

Minor moraines commonly trend parallel to the nearest down-glacier end moraine.

Where there are multiple sets of minor moraines with varying orientations, at least one set commonly lies parallel to the end moraine immediately down-glacier (Fig. 3.10a). In

a b

Figure 3.10: Minor moraines near the Bemis moraine. The black line is the distal edge of the moraine. a. Minor moraines parallel to the Bemis moraine on the western margin of the map area near Nemaha, Iowa. The glacier flow direction was to the west. b. Minor moraine sets, with trends marked by dashed black lines, at an oblique angle to the Bemis moraine west of West Des Moines, Iowa. some places, however, there are no sets of minor moraines related to the down-glacier end moraine (Fig. 3.10b). Additionally, end moraines sometimes truncate minor moraines immediately down-glacier from them, with a large angle between the minor moraine and end moraine orientations (Fig. 3.11).

Figure 3.11: Minor moraines, with trends marked by dashed black lines, truncated by the western edge of the West Bend moraine (black line) west of West Bend, Iowa. The eastern side of the moraine is its up-glacier side.

Along flow paths based on the most recent reconstruction of the Des Moines Lobe

(Hooyer & Iverson, 2002), rose diagrams indicating moraine-set orientations illustrate the overall tendency for multiple moraine sets, sometimes at high angles to one another

(up to 80°) (Fig. 3.12). Near the Bemis margin there is only one set of minor moraines that is parallel to the margin, albeit with one notable exception (Fig. 3.10).

Figure 3.12: Rose diagrams from four positions along a a. western, b. central, and c. eastern flow path, depicting the variability of minor moraine orientations.

Figure 3.12 continued:

Flow paths based on the Hooyer and Iverson (2002) reconstruction (Fig. 3.13) were used to compare the angular differences between minor moraine orientations and end moraines immediately down-glacier. End moraines that were considered include both the three major end moraines and more minor end moraine ridges, many of which are identified for the first time herein. As noted in Chapter 2, the goal was to explore whether end moraines generally represent surge-type advances. These advances should have an associated minor moraine set that reflects filling of transverse crevasses formed during surge-related longitudinal extension. Alternatively, some end moraines may be recessional, without associated minor moraine sets.

FP19

FP1 FP18

FP17 FP2

FP3 FP4 FP16

FP5 FP6 FP7 FP15

FP8

FP14 FP9 FP13 FP12 FP10 FP11

Figure 3.13: 19 flow paths across the Des Moines Lobe in Iowa along which orientations of minor moraines were determined. Blue dots indicate the center point of each sample area along the flow path.

The angular difference between the orientations of the nearest down-glacier end moraine and the sets of minor moraines was calculated at positions along each flow path

(Fig. 3.14). Where there are multiple minor moraine sets, the angular differences between the minor moraine sets both most parallel (mode min) and least parallel (mode max) to the end moraine immediately down-glacier were calculated and plotted against the distance along the flow path, with end moraine positions noted and the Des Moines

Lobe terminus at 0 km. Positions without data points do not have minor moraines or

have only one set of them; in that case the angular difference was considered a minimum angle (mode min).

In general, the three major end moraines have a set of minor moraines immediately up-glacier that is approximately parallel, 0-15o, to the end moraine. 15o is chosen as a threshold herein, based on the typical variability of orientations within a minor moraine set and the curvature of flow paths, both of which tend to increase angular differences.

For the Bemis moraine, nine of the 17 positions where there is a minor moraine up- glacier with no intervening end moraine have a minor moraine set approximately parallel to the Bemis moraine. At 11 of 16 such positions and 15 of 19 such positions for the

Altamont and Algona moraines, respectively, a minor moraine set parallels them immediately up-glacier. In general, each end moraine within the Des Moines Lobe had one set of minor moraine orientations approximately parallel to it along some of the flow paths considered. In a few instances, however, there were no minor moraine sets along any flow path that were parallel to the nearest down glacier end moraine (e.g., EM17, see Fig. 3.14).

Data in Figure 3.14 are summarized in Table 3.1, in which the minimum (mode min) and maximum angles (mode max) between minor moraine sets and end moraines directly down-glacier are tabulated and averaged among the flow paths for each end moraine. Averages for the minimum angles illustrate that, when multiple flow paths are available for an end moraine, angular differences larger than 15° are relatively rare and have large standard deviations. Averages for maximum angles illustrate that every end

moraine has at least one minor moraine set at a high angle to it (27-89°). The histograms of Figure 3.15 summarize the averages of Table 3.1.

A second way to summarize these data is to consider, not only the orientation of the minor moraine immediately up-glacier from an end moraine, but to consider the orientations of all minor moraine sets between the end moraine under consideration and the next end moraine up-glacier. Computing an average that includes all such minor moraines along all flow paths yields the average angular differences and standard deviations of Table 3.2. Although average minimum angular differences tend to be slightly larger than those of Table 3.1, angular differences from end moraines that include data from many flow paths tend to be below or at least near 15°. Larger values, particularly for the Bemis and Altamont moraines (22° and 23°), may reflect the curvature of flow paths and resultant non-parallelism that should increase with increasing distance up-glacier from the end moraines.

Figure 3.14: Minimum angular difference ( mode min) and maximum angular difference (mode max) plotted along 19 flow paths. Each end moraine intersected is labeled, with the terminal moraine located at zero kilometers. Missing data points are from areas without a second set of minor moraines or without any minor moraines.

Figure 3.14 continued:

Table 3.1: Minimum (top) and maximum (bottom) angular differences for each flow path (fp) between the trends of end moraines and orientations of minor moraine sets immediately up-glacier. Averages and standard deviations for each end moraine are also listed. Note that the blank spaces indicate that the flow path does not cross that moraine.

Mode Min - Immediately Up Glacier from End Moraine Flow Paths (west to east) fp1 fp2 fp3 fp4 fp5 fp6 fp7 fp8 fp9 fp10 fp11 fp12 fp13 fp14 fp15 fp16 fp17 fp18 fp19 Ave std. dev Bemis 0 7 3 2 8 17 36 23 25 54 13 21 9 10 27 4 49 18.11765 15.60122 Altamont 13 15 7 1 23 21 13 13 4 2 1 15 5 25 46 52 16 14.51723 Algona 17 15 16 13 8 7 1 7 5 5 7 14 2 6 16 12 16 11 7 9.736842 4.929419 Clare 16 15 16 73 1 6 2 6 6 6 12 6 0 5 12.14286 17.63866 End Moraines End Renwick 9 15 6 14 18 18 24 11 5 11 17 7 31 14.30769 7.1617 West Bend 15 8 4 13 11 1 16 3 17 52 14 13.76227 EM1 8 4 37 5 12 13.2 12.22129 EM2 16 25 32 70 10 5 7 11 22 22 18.9385 EM3 23 24 19 2 5 9 0 23 28 9 18 6 1 26 13.78571 9.806547 EM4 20 18 5 9 3 0 3 14 14 22 86 28 5 12 42 5 30 18.58824 20.12917 EM5 18 12 5 20 13.75 5.847008 EM8 8 12 10 10 1.632993 EM9 9 44 1 7 12 12 14 31 6 14 84 21.27273 23.00629 EM10 15 27 21 6 EM11 7 6 32 15 12.02775 EM12 3 5 19 9 7.118052 EM13 6 3 4.5 1.5 EM14 15 15 0 EM15 55 55 0 EM16 9 9 0 EM17 28 5 10 11 27 31 18.66667 10.24153 EM18 2 3 5 6 6 4 10 5.142857 2.415934 EM19 19 19 0 EM20 1 10 3 10 7 6.2 3.655133 EM21 11 11 0 EM22 0 0 0 EM23 3 11 16 10 5.354126 EM24 10 10 0 EM25 40 40 0 EM27 32 32 0 EM28 15 15 0 EM29 22 22 0 EM31 31 31 0

Table 3.1 continued:

Mode Max - Immediately Up Glacier from End Moraine Flow Paths (west to east) fp1 fp2 fp3 fp4 fp5 fp6 fp7 fp8 fp9 fp10 fp11 fp12 fp13 fp14 fp15 fp16 fp17 fp18 fp19 Ave std. dev Bemis 22 52 37 39 77 80 70 71 38 24 30 38 80 12 74 49.6 22.84674 Altamont 46 39 85 29 50 14 21 31 44 42 13 53 77 70 43.85714 21.34699 Algona 46 35 40 32 43 49 52 49 61 61 48 70 75 79 79 68 67 68 68 57.36842 14.33145 Clare 79 82 74 90 61 45 58 62 68 10 16 29 84 34 56.57143 24.96201 End Moraines End Renwick 45 31 52 31 73 74 80 34 22 57 78 79 60 41 54.07143 19.674 West Bend 37 37 42 38 63 11 28 36 73 34 39.9 16.37345 EM1 86 88 42 72 21.22891 EM2 78 81 81 87 36 74 87 28 69 21.82888 EM3 42 27 31 49 51 54 46 73 68 71 78 39 44 83 54 17.08801 EM4 31 27 50 37 49 53 51 69 70 78 88 68 39 27 82 68 48 55 18.724 EM5 87 61 68 49 66.25 13.77271 EM8 36 37 32 35 2.160247 EM9 38 89 52 51 62 62 64 82 89 37 62.6 18.11187 EM10 36 43 39.5 3.5 EM11 58 33 43 44.66667 10.27402 EM12 31 22 57 36.66667 14.83989 EM13 33 21 27 6 EM14 30 30 0 EM15 83 83 0 EM16 37 37 0 EM17 51 88 80 56 86 72.2 15.57434 EM18 21 26 21 57 34 31.8 13.467 EM19 70 70 0 EM20 84 12 27 30 38.25 27.2798 EM21 23 23 0 EM22 68 68 0 EM23 42 25 30 32.33333 7.133645 EM24 29 29 0 EM25 89 89 0 EM27 47 47 0 EM28 42 42 0 EM29 31 31 0 EM31 49 49 0

a b

Figure 3.15: Histograms of a. the average minimum (mode min) and b. maximum (mode max) angular differences for the end moraines studied.

Table 3.2: Averages and standard deviations like those listed in Table 1 but considering the orientations of all minor moraine sets between the end moraine under consideration and the next end moraine up-glacier.

Mode Min Mode Max End Moraine Average Standard Number of End Moraine Average Standard Number of Angular Deviation Sample Angular Deviation Sample Difference Positions Difference Positions

Bemis 22.32727273 15.15862136 55 Bemis 46.64102564 21.72495573 39 Altamont 23.27380952 20.94684383 84 Altamont 46.98734177 25.46006735 79 Algona 11.57251908 8.371188808 131 Algona 52.40495868 15.53964725 121 Clare 11.48 14.60169853 25 Clare 63.24 22.41299623 25 Renwick 15.05555556 8.51451194 18 Renwick 54.16666667 21.54646143 18 West Bend 14.94736842 11.4729599 19 West Bend 36.5625 15.18621394 16 EM1 11.625 9.986709919 8 EM1 66.66666667 19.88019673 6 EM2 15.14285714 17.38988681 14 EM2 59.78571429 24.26984776 14 EM3 14.72222222 9.18214182 18 EM3 54.77777778 16.33182155 18 EM4 16.69565217 18.64178286 23 EM4 54.82608696 18.20093685 23 EM5 13.4 5.953150426 15 EM5 56.85714286 12.47200948 14 EM8 10 1.632993162 3 EM8 35 2.160246899 3 EM9 23 27.41350032 16 EM9 60.33333333 17.33653817 12 EM10 16.4 8.731551981 10 EM10 41.75 14.11338018 8 EM11 10.33333333 9.877021593 6 EM11 48.83333333 9.753916592 6 EM12 9.857142857 8.043250434 7 EM12 42.28571429 14.17960594 7 EM13 6.157894737 5.96713065 19 EM13 31.76470588 16.47646954 17 EM14 16.66666667 2.357022604 3 EM14 38.66666667 6.12825877 3 EM15 52.33333333 1.885618083 3 EM15 75.33333333 5.436502143 3 EM16 11.66666667 3.091206165 3 EM16 31.66666667 4.496912521 3 EM17 25.33333333 11.63328558 9 EM17 74.14285714 14.02475945 7 EM18 8.642857143 4.185129164 14 EM18 59.54545455 26.96247775 11 EM19 25.625 15.70778072 8 EM19 58.57142857 7.228063223 7 EM20 8.363636364 5.756778254 11 EM20 36.625 27.98632479 8 EM21 8 3 4 EM21 52.25 17.25362281 4 EM22 0 0 1 EM22 68 0 1 EM23 9.555555556 4.597369081 9 EM23 29.85714286 8.790068607 7 EM24 10 0 1 EM24 29 0 1 EM25 43 3 2 EM25 89 0 1 EM27 32 0 1 EM27 47 0 1 EM28 15 0 1 EM28 42 0 1 EM29 25 3 4 EM29 44.33333333 9.428090416 3 EM31 33.77777778 7.799968344 9 EM31 44.77777778 4.184037863 9

CHAPTER 4: DISCUSSION

A. Evidence of stagnation

There are two contrasting hypotheses of the Des Moines Lobe advance and retreat.

The leading hypothesis is that the lobe underwent surge-like advances, to build its three major end moraines, with stagnation after each advance (Clayton et al., 1985; Kemmis,

1991). Alternatively, the lobe may have experienced active retreat (Gwynne, 1942;

Gwynne, 1951; Lawrence & Elson, 1953), such that negative mass balance and recession were accompanied by active ice flow. The results of this project indicate that 90% of the

Des Moines Lobe landscape in Iowa, not considering modern drainages, comprises stagnation landforms, including minor moraines, hummocky moraines, and ice-walled lake plains (Fig. 3.1). The prevalence of stagnation landforms across most of the lobe’s area suggests that the lobe underwent surge-like advances and stagnation.

Minor moraines cover 60% of the Des Moines Lobe footprint (Fig. 3.1), and diverse data indicate that they formed by filling of crevasses that extended to the bed.

Ankerstjerne (2010) studied a cross-section through a minor moraine in Story County,

Iowa, and concluded that the moraine formed by crevasse filling. Preconsolidation and density data indicated that the till of the moraines had been deposited under the glacier, rather than at its margin. Moreover, till fabrics based on anisotropy of magnetic susceptibility (e.g., Shumway and Iverson, 2009) indicated upward shear in the proximal side of the moraine, with longitudinal shortening and upward extension just down- glacier from the moraine crest where till was squeezed upward into the crevasse (Fig.

4.1). Earlier less comprehensive sedimentological work but on several minor moraines

Figure 4.1: Conceptual model of minor moraine formation due to filling of crevasse that extends to the bed (from Ankerstjerne, 2010) in Story County led Stewart et al. (1988) to also conclude the moraines were crevasse- fill ridges. Cline (2011) studied the geometric attributes of the minor moraines of Story

County and found that they are most commonly both symmetrical parallel to ice flow and periodically spaced consistent with crevasse-fill moraines associated with surge-type glaciers (Sharp, 1985b; Herzfeld & Mayer, 1997; Evans et al., 1999; Herzfeld et al.,

2004; Molnia, 2004; Schomacker & Kjær, 2007; Evans et al., 2009; Johnson et al.,

2010).

Crevasse-fill ridges are uniquely associated with the forefields of surge-type glaciers and are thought to form at the end of the surge phase during stagnation (e.g., Sharp,

1985a; Rea and Evans, 2011). Thus, the abundance of minor moraines within the area of

Des Moines Lobe reinforces the conclusion that the lobe underwent surge-like advances followed by stagnation.

Hummocky moraines develop in stagnant ice environments as ice-cored moraines begin to down-waste (Benn, 1992). Hummocky moraine development associated with surge-type glaciers is well documented; after the glacier surges, stagnant debris-laden ice at the glacier margin down-wastes (Kjær & Krüger, 2001; Evans & Rae, 2003;

Schomacker & Kjær, 2007). 31% of the footprint of the lobe is characterized by hummocky moraines. The Altamont moraine (Fig. 2.3, 4.2) is one of the best examples within the Des Moines Lobe. Hummocky end moraines are thus additional evidence of stagnation, a point also emphasized by others (e.g., Kemmis, 1991).

Figure 4.2: Hummocky topography of the Altamont moraine, north of Boone, Iowa, with flow direction indicated.

Newly mapped ridges, 5-10 m high, with a width of ~1 km. (Fig. 2.4b, c) are assumed to be end moraines. This assumption is based on the lateral continuity of the

ridge and its sub-parallel relationship to the previously identified end moraines. These ridges likely formed by processes similar to those that formed the broad hummocky end moraines. Differences in geometry between these two styles of end moraines may be explained by the number of times the glacier advanced to a given location. Low amplitude, ~5 m high, ridges may reflect a single advance, whereas broad hummocky moraines with till thicknesses of ~10-15 m may reflect multiple advances to approximately the same position (Kjær & Krüger, 2001; Evans & Rea, 2003;

Schomacker & Kjær, 2007). Alternatively, lower amplitude ridges may reflect a margin position that did not persist for an extended period. Although the origin of these ridges is uncertain, their morphology suggests that they likely formed along the surging margin of the glacier and can thus be considered end moraines.

Although less prevalent than minor moraines or hummocky end moraines, ice-walled lake plains cover 6% of the map area and also develop in stagnant ice environments.

Supraglacial lakes develop as debris and melt water accumulate in depressions on the stagnant ice surface. The sediment collected in the lakes is deposited on the ground surface when the stagnant ice melts (Ham & Attig, 1996; Johnson & Clayton, 2003;

Clayton et al., 2008). Ice-walled lake plains, therefore, also indicate that the Des

Moines Lobe was characterized by stagnant ice.

The observation that 90% of the Des Moines Lobe landscape in Iowa consists of stagnation landforms does not provide information regarding the number of surges or relative timing of stagnation over different parts of the map area. Moreover, newly identified end moraine ridges without stagnation topography leave open the possibility

that that parts of the lobe’s margin left recessional moraines formed during active retreat.

Minor moraine orientations can provide relevant information on these issues.

B. Significance of minor moraine orientations

Nearly all of the end moraines studied—the three major hummocky end moraines, other previously mapped end moraines, and the 25 newly mapped end moraine ridges— have a set of moraine ridges immediately up-glacier that are parallel or sub-parallel to them. Given that minor moraine ridges are likely transverse crevasse-fill ridges that formed immediately after surge-like advances, most end moraine ridges can be interpreted as being the result of surge-like advances. Thus, the likelihood that there were many more than three surge-like advances seems high. This conclusion is supported by the observation that in most areas and for every end moraine there is more than one set of minor moraines, with some of those sets having orientations that deviate greatly from the orientation of the nearest end moraine down-glacier. Minor moraines at a significant angle to the adjacent end moraine down-glacier correspond to end moraines farther down-glacier and have been presumably overridden by a later surge. This sort of re-advance over previous surge-type topography is common in modern surge-type glacier forefields (Kjær & Krüger, 2001; Evans & Rea, 2003; Schomacker & Kjær,

2007) and supported by the observation that minor moraines are sometimes truncated by end moraines (Fig. 3.11). Surge-like re-advance is also supported by the superposition of minor moraines on ice-walled lake plains (Fig. 3.5). The alternative—that most end moraine ridges are recessional and that glacier flow during recession was in the same direction as surge motion—seems unlikely given the diversity of flow directions

indicated by minor moraine orientations. Moreover, the spatial prevalence of stagnation features, together with their tendency to be focused relatively near modern glacier margins, argues for multiple surge events to multiple locations to progressively build stagnation landforms spread over 90% of the lobe’s landscape.

Minor moraines have previously been used to reconstruct flow paths and the ice surface of the Des Moines Lobe (Clark, 1992; Hooyer & Iverson, 2002).

Reconstructions have involved making the assumption that minor moraines were perpendicular to flow when the lobe was at its maximum extent building the Bemis moraine, so that all of the minor moraines formed during the Bemis phase. However, multiple minor moraine orientations within the same area suggest that minor moraines did not form during the Bemis phase and that different sets reflect separate surge-like advances of the lobe. Thus, the extent to which minor moraines, particularly those moraines well up-glacier from the Bemis moraine, can be used to infer flow direction during the lobe’s maximum extent is uncertain. Minor moraines far up-glacier that are best preserved and hence chosen for reconstructions are most likely to be associated with end moraines that are well up-glacier from the Bemis moraine. In contrast, far up- glacier from the Bemis moraine, minor moraines associated with the Bemis moraine advance are likely to be either poorly preserved or destroyed due to multiple overriding episodes. Also possible is that transverse crevassing required for minor moraine formation did not extend many tens of kilometers up-glacier from the glacier margin, such that all moraines far up-glacier from the Bemis moraine reflect later advances.

Given the strong likelihood that most minor moraines well up-glacier from the Bemis

moraine are associated with later advances of the lobe, a useful exercise would be to assess the sensitively of Des Moines Lobe reconstructions to disregarding minor moraine orientations that extend a threshold distance up-glacier from the Bemis moraine.

C. Lack of flow-parallel features

Flow-parallel features such as drumlins or flutes are characteristic of some surge-type glaciers (Evans et al., 2009; Johnson et al., 2010). Flow-parallel features are rare but present within the footprint of the lobe. Some of these features connect segments of minor moraines and are best interpreted as flow-parallel crevasse fills (Fig. 4.3). During surges, crevasses commonly form at roughly right angles, with one set parallel to flow

(Fig. 4.4). Other rare flow-parallel features are less discrete lineations, extending ~500 m and superimposed across multiple washboard moraines (Fig. 4.5). These could also be a consequence of longitudinal crevasse fills, although their poor continuity along their lengths makes that interpretation uncertain. Drumlins and flutes are absent in the map area, and although one or both of these landforms can be absent from the forefields of modern surge-type glaciers (e.g. Loken, 1969; Holdsworth, 1977; Roberts et al., 2009;

Evans et al., 2010), their absence from the Des Moines landscape is puzzling.

A leading hypothesis for drumlin formation requires a partially frozen bed

(Whittecar & Mickelson, 1979; Stanford & Mickelson, 1985; Colgan & Mickelson,

1997; Hooke & Medford, 2013). As a glacier flows over partly thawed permafrost, thawed zones are eroded, with some sediment deposited at the crests and lee sides of the frozen zones (Stanford & Mickelson, 1985; Hooke & Medford, 2013). The Des Moines

Figure 4.3: Longitudinal moraines (red dashed lines) interpreted as longitudinal crevasse fills. Yellow dashed lines indicate minor moraines (from Cline, 2011).

Figure 4.4: Photograph of the surge-type Bering Glacier, Alaska, in 1993 (USGS) showing prominent flow-parallel crevasses.

Figure 4.5: Area with flow-parallel ridges northeast of Nevada, Iowa.

Lobe did not have a frozen bed as it advanced out of balance with the climate into Iowa

(Clark, 1994; Hooyer & Iverson, 2002; Mickelson & Colgan, 2003), so if this hypothesis is correct it could explain the absence of drumlins in Iowa and their ubiquity beneath some more northerly Midwest lobes of the Laurentide ice sheet (Whittecar & Mickelson,

1979; Stanford & Mickelson, 1985; Sharp, 1988; Colgan & Mickelson, 1997). On the other hand, several other popular hypotheses for drumlin formation do not require a frozen or partly frozen bed (Boulton, 1987; Hindmarsh, 1998; Fowler, 2009)

Flutes, a few decimeters to a few meters high and hundreds of meters long, are a common flow-parallel landform in the forefields of modern surge-type glaciers (Evans &

Rea, 2003). Their absence from the Des Moines Lobe landscape may reflect their tendency to be readily eroded by post-glacial processes (Benn & Evans, 2010).

D. Modern analog

The multiple advances of the Des Moines Lobe—probably well in excess of the three surge-like advances into Iowa that have been typically postulated—invites comparisons with modern surge-type glaciers that repeatedly and periodically re-advance across their forefields (Evans et al., 1999; Schomacker & Kjær, 2007; Fleisher et al., 2010).

Although there is no evidence for periodic advance of the lobe and thus describing it as a surge-type glacier would be overreaching, as noted in Chapter 1, many of the lobe’s characteristics invite comparisons with surge-type glaciers and identification of a modern analog.

The Bering Glacier (Fig. 4.4), the largest surge-type glacier in North America (5200 km2 and 190 km long), is perhaps the best analog, although imperfect. During its last surge from 1993 to 1995 it advanced at a rate of 365-2700 m yr-1, thus spanning the

2000 m yr-1 advance rate of the Des Moines Lobe determined by Clayton et al. (1985) using the radiocarbon chronology of the lobe. The Bering Glacier rests on a soft bed and during its last surge was under high basal water pressure (Fleisher et al., 2010; Molnia &

Post, 2010) similar to the inferred basal dynamics of the Des Moines Lobe’s advance into Iowa (Fig. 1.4, Clark, 1994; Hooyer & Iverson, 2002). Transverse crevassing was pervasive on the surface of the Bering Glacier during its last surge (Fig. 4.6a). These crevasses locally extended to the bed and produced moraines remarkably similar in shape and scale to the minor moraines of the Des Moines Lobe (Fig. 4.6c). Sudden release of pressurized water from the bed at the end of the Bering Glacier surge locally resulted in scoured channels and eskers, features also associated with Des Moines Lobe

(Fleisher et al., 2010). In other ways the comparison breaks down. Bering Glacier is confined by valley walls along most of its length, and over some of its margin, terminates in a large lake. End moraine ridges are not as prominent as in the Des Moines

Lobe footprint. Moreover, some of the Bering Glacier forefield is mantled by flutes and poorly developed drumlins (Fleisher et al., 2010).

Figure 4.6: a. Transverse crevasses formed during the 1993-1995 surge of the Bering Glacier, Alaska (photo by Bernhard Edmaier; Cline, 2011). b. Aerial photograph of a ~4 km2 area exposed by the retreat of the Bering Glacier following the 1993-1995 surge (Molnia, 2004). c. Minor moraines of the Des Moines Lobe at about the same scale as in b. (from Cline 2011).

CHAPTER 5: CONCLUSIONS

Stagnation landforms covering 90% of the Des Moines Lobe’s footprint in Iowa, including minor moraines (60%), hummocky topography (31%), ice-walled lake plains

(6%) and doubly blown-out donuts (3%), indicate that the glacier underwent widespread stagnation at multiple times. The extent of minor moraines, only associated with surge- type glaciers (e.g., Sharp, 1985), suggests that episodes of stagnation were preceded by surge-like advances of the lobe.

Multiple orientations of minor moraines at most locations indicate that most of the lobe’s landscape reflects the action of two or more advances. Of the approximately 30 end moraines mapped, most of them, over at least part of their lengths, have a parallel

(<15o) set of minor moraines indicating that most moraines are likely the result of a surge-like advances. Additionally, all moraines have at least one secondary set of minor moraines at a high angle to the margin. These observations, considered collectively, indicate that there were multiple surge-like advances of the Des Moines Lobe—likely many more than the three advances thought to have formed the Bemis, Altamont, and

Algona moraines (Clayton et al., 1985; Kemmis, 1991). Minor moraines that are both far from the Bemis moraine and well preserved are probably associated with later advances of the lobe and should not be used as flow direction indicators when the lobe was at its maximum extent building the Bemis moraine.

The surge-type, Bering Glacier, is a good, albeit imperfect, modern analog for the

Des Moines Lobe, with speeds during surges, basal conditions, and crevasse-fill ridges

similar to those inferred for the lobe. Unlike the Bering Glacier, however, the Des

Moines Lobe in Iowa was not confined along it lateral margins and left no flutes that have been preserved or drumlins.

REFERENCES

Alley, R.B., 1991. Deforming-bed origin for southern Laurentide till sheets? Journal of Glaciology, 37(125), pp. 67-76.

Alley, R.B., 1989. Water-Pressure Coupling of Sliding and Bed Deformation: 1. Water system. Journal of Glaciology, 119, pp. 108-18.

Alley, R.B., Blankenship, D.D., Bentley, C.R. & Rooney, S.T., 1987. Till beneath ice stream B: 3. Till deformation evidence and implications. Journal of Geophysical Research: Solid Earth, 92, pp.8921-8929. Ankerstjerne, S. M., 2010. Origin of the 610th Avenue Moraine of the Des Moines Lobe, Story County, Iowa. (unpublished M.S. thesis, Iowa State University)

Baker, R.G., 1996. Pollen and plant macrofossils. In Bettis, E.A., Quade, D.J., & Kemmis, T.J., eds. Hogs, bogs, and logs: Quaternary deposits and environmental geology of the des Moines Lobe. Iowa Department of Natural Resources, Guidebook Series 6, pp.105-109.

Benn, D.I., 1992. The genesis and significance of ‘hummocky moraine’: evidence from the Isle of Skye, Scotland. Quaternary Science Reviews, 11, pp. 781-99.

Benn, D.I & Evans, D.J.A., 2010. Glaciers and Glaciation. London: Hodder.

Bettis, E.A., Quade, D.J. & Kemmis, T.J., 1996. Hogs, Bogs, & Logs: Quaternary deposits and environmental geology of the Des Moines Lobe. Geological Survey Bureau Guidebook Series 18.

Boone, S.J. & Eyles, N., 2001. Geotechnical model for great plains hummocky moraine formed by till deformation below stagnant ice. Geomorphology, 38, pp. 109-24.

Boulton, G.S., 1996. Theory of glacial erosion, transport and deposition as a consequence of subglacial sediment deformation. Journal of Glaciology, 42(140), pp. 43-62.

Boulton, G.S., 1987. A theory of drumlin formation by subglacial deformation. In J. Menzies and J. Rose, eds. Drumlin Symposium. Rotterdam: Balkema, pp. 25–80.

Burras, C.L. & Scholtes, W.H., 1987. Basin Properties and Postglacial Erosion Rates of Minor Moraines in Iowa. Soil Science Society of America Journal, 51(6), pp.1541-47.

Chamberlin, T.C., 1883. Preliminary paper on the terminal moraine of the second glacial epoch. U.S. Geological Survey Report, 3, pp. 291-402.

Clarke, G.K., 1987. Fast Glacier Flow: Ice Streams, Surging, and Tidewater Glaciers. Journal of Geophysical Research, 92 pp. 8835-8841.

Clark, P.U., 1994. Unstable Behavior of the Laurentide Ice Sheet over Deforming Sediment and its Implications for Climate Change. Quaternary Research, 41, pp. 19-25.

Clark, P.U., 1992. Surface form of the southern Laurentide Ice Sheet and its implications to ice-sheet dynamics. Geological Society of America Bulletin, 104, pp.242-40.

Clark, P.U., & Walder, J.S., 1994. Subglacial drainage, eskers, and deforming beds beneath the Laurentide and Eurasian ice sheets. Geological Society of America Bulletin, 106, pp. 304-14.

Clayton, L., Attig, J.W., Ham, N.R., Johnson, M.D., Jennings, C.E., & Syverson, K.M., 2008. Ice-walled-lake plains: Implications for the origin of hummocky glacial topography in middle North America. Geomorphology, 97, pp. 237-48.

Clayton, L., Attig, J.W., & Mickelson, D.M., 1999. Tunnel channels formed in Wisconsin during the last glaciation. Special Papers Geological Society of America, 337, pp. 69-82.

Clayton, L. & Freers, T. F., 1967. Glacial Geology of the Missouri Coteau and Adjacent Areas. North Dakota Geological Survey Miscellaneous Series 30.

Clayton, L. & Moran, S.R., 1982. Chronology of Late Wisconsinan Glaciation in Middle North America. Quaternary Science Reviews, 1, pp.55-82.

Clayton, L. Teller, J.T., & Attig, J.W., 1985. Surging of the southwestern part of the Laurentide Ice Sheet. Boreas, 14, pp. 235-41.

Cline, M. D., 2011. Spatial analysis of Des Moines Lobe washboard moraines using LiDAR data. (unpublished M.S. thesis, Iowa State Uiversity).

Cofaigh, C.O., 1996, Tunnel valley genesis. Progress in Physical Geography, 20(1), pp. 1-19.

Colgan, P.M., & Mickelson, D.M., 1997. Genesis of streamlined landforms and flow history of the Green Bay Lobe, Wisconsin, USA. Sedimentary Geology, 111, pp.7-25.

Cracknell, A. P. & Hayes, L., 1991. Introduction to Remote Sensing. London: Taylor and Francis.

Cuffey, K.M. & Paterson, W.S., 2010. The Physics of Glaciers. Oxford: Butterworth- Heinemann.

Elachi, C. & van Zyl, J., 2006. Introduction to the Physics and Techniques of Remote Sensing (2nd ed.). Hoboken, NJ: John Wiley & Sons.

Elson, J.A., 1968. Washboard moraines and other minor moraine types. In R.W. Fairbridge, ed. The Encyclopedia of Geomorphology. Reinhold Book Corp. pp.1213-19.

Eyles N., Boyce, J.I. & Barendregt, R.W., 1999. Hummocky moraine: sedimentary record of stagnant Laurentide Ice Sheets resting on soft beds. Sedimentary Geology, 123, pp.163-74.

Evans, D.J.A., Lemmen, D.S., & Rea, B.R., 1999. Glacial landsystems of the southwest Laurentide Ice Sheet: modern Icelandic analogues. Journal of Quaternary Science, 14, pp. 673-91.

Evans, D.J.A., Nelson, C.D. & Webb, C., 2010. An assessment of fluting and “till esker” formation on the foreland of Sandfellsjökull, Iceland. Geomorphology. 114, pp.453-65.

Evans, D.J.A., & Rea, B.R., 2003. Surging glacier landsystem. In D.J.A. Evans, ed. Glacial Landsystems. London: Arnold. pp. 259-88.

Evans, D.J.A., Twigg, D.R, Rea, B.R., & Orton, C., 2009. Surging glacier landsystem of Tungnaárjökull, Iceland. Journal of Maps, pp. 134-`51.

Fleisher, P.J., Bailey, P.K., Natel, E.M., Muller, E.H., Cadwell, D.H. & Russell, A., 2010. The 1993-1995 surge and foreland modification, Bering Glacier, Alaska. The Geological Society of America Special Paper 462. pp.193-216.

Fowler, A.C., 2009. Instability modelling of drumlin formation incorporating lee-side cavity growth. Proceedings of the Royal Society, 465, pp. 2681-2702.

Gwynne, C.S., 1942. Swell and swale pattern of the Mankato Lobe of the Wisconsin Drift Plain in Iowa. The Journal of Geology, 50(2), pp.200-08.

Gwynne, C.S., 1951. Minor moraines in South Dakota and Minnesota. Geological Society of America Bulletin, 62, pp.233-50.

Hallberg, G. R., Lineback, J. A., Mickelson, D. M., Knox, J. C., Goebel, J. E., Hobbs, H. C., … Dreeszen, V. H., 1991. Quaternary Geologic Map of the Des Moines 4o x 6o Quadrangle, United States. U.S. Geological Survey.

Ham, N.R. & Attig, J.W., 1996. Ice wastage and landscape evolution along the southern margin of the Laurentide Ice Sheet, north-central Wisconsin. Boreas, 25(3), pp. 171-86.

Herzfeld, U.C. & Mayer, H., 1997. Surge of Bering Glacier and Bagley Ice Field, Alaska: an update to August 1995 and an interpretation of brittle deformation patterns. Journal of Glaciology, 43 (145), pp.427-34.

Herzfeld U.C., Clarke, G.K.C., Mayer, H. & Greve, R., 2004. Derivation of deformation characteristics in fast-moving glaciers. Computers & Geosciences, 30, pp.291- 302.

Hindmarsh, R.C.A., 1998. The stability of a viscous till sheet coupled with ice flow, considered at wavelengths less than the ice thickness. Journal of Glaciology, 44, pp. 285-92.

Holdsworth, G., 1977. Surge activity on the Barnes Ice Cap. Nature, 269, pp. 588-90.

Hooke, R.L. & Medford, A., 2013. Are drumlins a product of thermo-mechanical instability? Quaternary Research, 79, pp. 458-64.

Hooyer, T.S. & Iverson, N.R., 2002. Flow mechanism of the Des Moines lobe of the Laurentide ice sheet. Journal of Glaciology, 48(163), pp.575-85.

Illinois State Geological Survey, 2008. Quaternary glaciations in Illinois: Illinois State Geological Survey, GeoNote 3, available online at www.isgs.uiuc.edu (accessed October 2013).

Iowa Department of Natural Resources. LiDAR. available online at www.iowadnr.gov (accessed October 2013).

Jennings, C.E., 2006. Terrestrial ice streams—a view from the lobe. Geomorphology, 75, pp. 100-24.

Johnson, M.D., & Clayton, L., 2003. Supraglacial landsystems in lowland terrain. In D.J.A. Evans, ed. Glacial Landsystems. London: Arnold. pp. 228-58.

Johnson, M.D., Schomacker, A., Benediktsson, Í.Ö., Geiger, A.J., Ferguson, A., & Ingólfsson, Ó., 2010. Active drumlin field revealed at the margin of Múlajökull, Iceland: A surge-type glacier. Geology, 38, pp. 943-46.

Kamb, B., Raymond, C.F., Harrison, W.D., Englehardt, H., Echelmeyer, K.A., Humphrey, N., Brugman, M.M. & Pfeffer, T., 1985. Glacier surge mechanism: 1982-1983 surege of Variegated Glacier, Alaska. Science, 227, pp.469-79.

Kemmis, T.J., Hallberg, G.R. & Lutenegger, A.J., 1981. Depositional environments of glacial sediments and landforms on the Des Moines lobe, Iowa. Iowa Geological Survey Guidebook Series 6.

Kemmis, T.J., 1991. Glacial Landforms, Sedimentology, and Depositional Environments of the Des Moines Lobe, Northern Iowa. (unpublished Ph.D. thesis, University of Iowa)

Kjær, K.H., & Krüger, J., 2001. The final phase of dead-ice moraine development: processes and sediment architecture, Kötlujökull, Iceland. Sedimnetology, 48, pp. 935-52.

Lawrence, D.B. & Elson, J.A., 1953. Periodicity of deglaciation in North America since the Late Wisconsin Maximum. Geografiska Annaler, 35(2), pp.83-104.

Leverett, F., 1932. Quaternary geology of Minnesota and parts of adjacent states. Geological Survey Professional Paper 161.

Loken, O.H., 1969, Evidence os surges on the Barnes Ice Cap, Baffin Island. Canadian Journal of Earth Sciences, 6, pp. 899-901.

Macbride, T.H., 1909, Geology of Hamilton and Wright counties. Iowa Geological Survey Annual Report, 20, pp. 97-149.

Mathews, W.H., 1974. Surface profiles of the Laurentide Ice Sheet in its marginal areas. Journal of Glaciology, 13(7), pp.37-43.

Measures, R.M., 1984. Laser Remote Sensing: Fundamentals and Applications. New York: John Wiley & Sons.

Mickelson, D.M., & Colgan, P.M., 2003. The southern Laurentide Ice Sheet. Development in Quaternary Science, 1, pp.1-16.

Molnia, B.F., 2004. Glossary of Glacier Terminology: A glossary providing the vocabulary necessary to understand the modern glacier environment. USGS Open-file report 2004-1216.

Molnia, B.F. & Post, A., 2010. Surges of the Bering Glacier. The Geological Society of America Special Paper 462. pp. 291-316.

Patterson, C.J., 1998. Laurentide glacial landscapes: The role of ice streams. Geology, 26(7), pp. 643-46.

Patterson, C.J., 1997. Southern Laurentide ice lobes were created by ice streams: Des Moines Lobe in Minnesota, USA. Sedimentary Geology, 111, pp.249-261.

Prest, V.K., 1968. Nomenclature of moraines and ice-flow features as applied to the glacial map of Canada. Geological Survey of Canada, Paper 67-57.

Quade, D.J., Giglierano, J.D., & Bettis, E.A., 2005. Surficial geologic map of the Des Moines Lobe of Iowa phase 7: Clay and Osceola Counties. Iowa Geological Survey.

Quade D.J. & Seigley, L.S., 2006. Quaternary Geology of the Storm Lake Area, Iowa: Storm Lake Outlet, Des Moines Lobe Moraines, Kames, Valley Trains, Minor Moraines and Tazewell Till Plain. Geological Society of Iowa Guidebook 78.

Rea, B.R. & Evans, D.J.A., 2011. An assessment of surge-induced crevassing and the formation of crevasse squeeze ridges. Journal of Geophysical Research: Earth Surface, 116, pp. 1-17.

Roberts, D.H., Yde, J.C., Knudsen, N.T., Long, A.J. & Lloyd, J.M., 2009. Ice marginal dynamics during surge activity, Kuannersuit Glacier, Disko Island West Greenland. Quaternary Science Reviews, 28, pp. 209-22.

Roush, J.J., Lingle, C.S., Guritz, R.M., Fatland, D.R. & Voronina, V.A., 2003. Surge- front propagation and velocities during the early 1993-95 surge of Bering Glacier, Alaska, U.S.A., from sequential SAR imagery. Annals of Glaciology, 36, pp.37-44.

Ruhe, R.V., 1969. Quaternary Landscapes in Iowa. Iowa State University Press, Ames.

Ruhe, R.V., 1952. Topographic Discontinuities of the Des Moines Lobe. American Journal of Science, 250, pp. 46-56

Schomacker, A., & Kjær, K.H., 2007. Origin and de-icing of multiple generations of ice- cored moraines at Brúarjökull, Iceland.

Schwert. D.P., & Torpen, H.J., 1996. Insect remains: a faceted eye’s perspective on the advance of the Des Moines Lobe into north-central Iowa. In Bettis, E.A., Quade, D.J., & Kemmis, T.J., eds. Hogs, bogs, and logs: Quaternary deposits and environmental geology of the des Moines Lobe. Iowa Department of Natural Resources, Guidebook Series 6, pp. 99-104.

Sharp, D.R., 1988. Late glacial landforms of Wollaston Peninsula, Victoria Island, Northwest Territories: product of ice-marginal retreat, surge, and mass stagnation. Canadian Journal of Earth Science, 25, pp. 262-79.

Sharp, M., 1985a. “Crevasse-fill” ridges – A landform type characteristic of surging glaciers? Geografiska Annaler, 67A(3/4), pp.213-20.

Sharp, M., 1985b. Sedimentation and stratigraphy at Eyjabakk ajökull—An Icelandic surging glacier. Quaternary Research, 24(3), pp. 268-284.

Shreve, R.L., 1985. Esker characteristics in terms of glacier physics, Katahdin esker system, Maine. Geological Society of America Bulletin, 96, pp. 639-46.

Shumway, J.R. & Iverson. N.R., 2009. Magnetic fabrics of the Douglas Till of the Superior lobe: exploring bed-deformation kinematics. Quaternary Science Reviews, 28, pp. 107-19.

Stanford, S.D. & Mickelson, D.M., 1985. Till fabric and deformational structures in drumlins near Waukesha, Wisconsin, U.S.A. Journal of Glaciology, 31, pp. 220- 28.

Stewart, R.A., Bryant, D. & Sweat, M.J., 1988. Nature and origin of corrugated ground moraine of the Des Moines Lobe, Story County, Iowa. Geomorphology, 1, pp.111-30.

Stokes, C.R. & Clark, C.D., 2001. Paleo-ice streams. Quaternary Science Reviews, 20, pp. 1437-57.

Stokes, C.R. & Clark, C.D., 1999. Geomporhpological criteria for identifying Pleistocene ice streams. Annals of Glaciology, 28, pp. 67-74.

Walder, J.S. & Fowler A., 1994. Channelized subglacial drainage over a deformable bed. Journal of Glaciology, 40(134), pp. 3-15.

Whittecar, G.R. & Mickelson, D.M., 1979. composition, internal structure, and a hypothesis of formation for drumlins, Waukesha County, Wisconsin, U.S.A. Journal of Glaciology, 22, pp. 357-71.